Fig. 1: Linear mixed effect models of group-by-time relationships with rate of cortical thickness change.
LME models described in parts A-D include age, age2, sex, and scanner as fixed effect covariates and a random subject-specific intercept. A F-test maps of group-by-time relationships with cortical thickness thresholded at p < 0.01 and p < 0.05 (uncorrected) indicate bilateral areas (described in the text) in which the rate of cortical thickness differs by clinical group. B 7 ROIs were retained for further analyses, by clustering areas from part A that pass the p < 0.01 threshold and are at least 100 mm2 in area. C T-tests (two-sided) across clusters (FDR corrected) to further assess relationships between cortical thickness change in each pair of groups indicate that cortical thinning occurs at a steeper rate in the CHR-C group, compared to the CHR-NC and HC groups in the clusters shown in blue in the image above. In the CHR-NC group, cortical thinning occurs at a steeper rate compared to the HC group in three clusters. D To further visualize rates of cortical thickness change by group, the average cortical thickness of left hemisphere clusters (left ROI) is plotted by time from baseline separately for each group. Plot indicates that on average, CHR-C participants have steeper cortical thinning compared to CHR-NC and HC participants.
